Voxtral trascrive alla velocità del suono.

Mistral AI has released Voxtral Transcribe 2, a new generation of speech-to-text models featuring state-of-the-art transcription quality and diarization. The release includes two models: Voxtral Mini Transcribe V2 for batch processing and Voxtral Realtime for live, low-latency applications like voice agents. Key features include support for 13 languages, contextual biasing for specific terminology, and word-level timestamps. Voxtral Realtime is notably distributed with open weights under an Apache 2.0 license, allowing for on-device deployment. The models are positioned as a highly efficient and cost-effective solution for various applications, including meeting intelligence, contact center automation, and media subtitling.